Read any story on the Brown-Mandel Senate race in Ohio lately, and you’ll inevitably see some commentary from Mandel himself or his campaign talking about how he’s gone from 13 points down to tying the race.
And if you looked at the trendlines recently (well, before the DNC Convention at least) on Talking Points Memo or Real Clear Politics, you’d think that Mandel might of had a point.
But like everything else Mandel has said in the campaign, his claim of coming back from a 13-point deficit to a tie is a lie. Period.
Let’s start with some basics about polling. One, it’s generally a good idea to look at the aggregate polling data to get a totality of the picture before making assumptions of the race. Two, not all pollsters are created equal. You can’t compare what one poll says to another and suggest there’s a trend line between the two. There’s a difference in sample size, questioning, timing, margins of error, etc. from poll to poll. And yet, that’s exactly what Mandel is trying to do when he compares the 12-point lead Brown had in July with Quinnipiac, a reputable pollster with a proven track record in Ohio to the more recent mail-in Columbus Dispatch poll, which has a national reputation of being a joke of a poll that uses a methodology so novel, no actual pollster uses it.
A fairer comparison would be to take a look at polling outfits trends over time since the last time they polled the race. If you do that, you’ll find that the race has tightened, but not nearly as much as Mandel would like you believe. Look, there is no polling outfit in the nation that is more widely regarded as having a polling bias in favor of Republicans than Rasmussen Reports. And its latest poll in the race in showed a four-point swing in Mandel’s favor from late May to mid-August.
Quinnipiac showed a five-point swing from late July to mid-August. Public Policy Polling released a poll today that shows Brown up by eight, which would still just be a two-point swing for Mandel since a month ago. There has been no poll in Ohio that is publicly available that suggests that Josh Mandel has managed a 13-point swing. None. So at best, Mandel’s bounce is +4, not +13. But Mandel wouldn’t be the first guy to exaggerate the size of his, um, poll numbers.
In fact, even with the polls showing a narrowing of the race, Brown still enjoys a lead near 50% and beyond the margin of error of all polls except Rasmussen, the University of Cincinnat’s Ohio Poll and the Dispatch poll. And not even Rasmussen, which has constantly polled the race, shows more than 4 points in Mandel’s favor lately (and it has not yet polled to see if Brown is seeing any post-Democratic convention bounce that Obama is showing in Ohio.)
But Rasmussen has a documented history of producing polls with a Republican bias. The U.C. Ohio Poll has traditionally been a reliable pollster, some had considered it the gold standard of polling in Ohio. However, this most recent poll had a sample size much smaller than other polls have used, and it has traditionally used in polling in Ohio. Furthermore, it’s own crosstabs concede that it polled so few Independents and African-Americans that its numbers on those demographics are unreliable. Translation: It oversampled Republicans.
We are not alone in criticizing the Dispatch‘s unconventional “mail in” response poll. If it had a proven track record, you’d have to wonder then why no other polling outfit in the nation (either privately for campaigns or in public polling) utilizes its methodology. For example, at this point in the 2010 gubernatorial election, the Dispatch‘s poll claimed that John Kasich had a twelve-point lead. The Strickland campaign protested and claimed their own internal polling showed a much closer race. Some political pundits, like Stuart Rothenberg, agreed that the Dispatch poll was entirely unreliable. Either Rothenberg was correct, or somehow John Kasich nearly blew a twelve-point lead in the final sixty days of the 2010 election.
In fact, in 2008, Nate Silver, of FiveThirtyEight.com fame (now with the New York Times) had the Dispatch poll dead last in his pollster rankings of reliable pollsters. It slightly improved in the 2010 rankings, but it was still ranked as less reliable than Research 2000, a polling outfit that would later collapse under allegations it was simply just manufacturing polling results. Yes, the Dispatch poll was ranked less reliable than an outfit accused of just flat out making up the numbers, folks.
Simply put, PPP and Quinnipiac have a much better proven track record in accurately predicting races. Yes, they show the race has narrowed, but that’s not surprising. First, Mandel continues to enjoy the support of millions in third parties ads that are almost all uniformally seeking to either tear Brown down or build Mandel up. Second, Mandel had virtually nowhere to go than up as a challenger. In other words, races always seem to tighten at this point in the cycle. Even though Rasmussen shows it tied, it also only shows no more than a +4 movement for Mandel. The notion that Mandel has closed +13 points with Brown recently just simply is not supported by the polling data.
Could Mandel still win? It’s possible. But political history of public opinion polling would strongly suggest that Brown is a safe bet for re-election. And keep in mind that starting next month, early voting begins and over half of the ballots in the election will be cast early. Mandel simply hasn’t closed the gap quickly enough to suggest that he can still win by the time most of the ballots will actually be cast.
The bright side of today’s PPP for Mandel is that Brown’s approval rating is slightly underwater at a 42% approval rating to 44% disapproval. But that still makes Brown the most popular Ohio politician compared to President Obama, Senator Rob Portman, Gov. John Kasich, or Treasurer Josh Mandel. The bad side? The same poll shows that Mandel’s favorability/unfavorability rating is 33%/44%. Eleven points underwater makes it hard for a challenger to catch up in a month against an incumbent that voters are split on his or her approval.
Also complicating Mandel’s chances is the fact that the same polling continues to show Obama leading Ohio by five. From Politico to Chuck Todd at MSNBC, there’s been considerable chatter that the insiders of the Romney campaign simply do not believe they can win Ohio. If Romney were to give up on Ohio, it would be devestating for Mandel’s Senate campaign.
In the end, Mandel has too little time, too little bounce, and is too underwater that it is likely he can still win this race. The safe bet is that Sherrod Brown wins re-election.